NLS issue

Renita · Post by **Renita** » Tue Jul 01, 2008 11:33 pm

Hi,

I am new to NLS and am having an issue accessing turkish characters.

I am using ISO8859-1 as my project default.

The source database is an SQL server from where I take the data transform the same and output it into an Oracle server.

When I say view data I get the data from the server, however I am unable to see the special turkish characters; instead get a ? displayed therein.

Can you please let me know if I am going wrong using NLS. Since I have never used NLS before I am unsure of the changes that need to be made.

Any inputs and advice on the same would be highly appreciated.

Thanks,
Renita

ArndW · Post by **ArndW** » Wed Jul 02, 2008 12:25 am

The problem might be that the "view data" window doesn't always show NLS characters correctly, the only way to be 100% certain that a problem is occurring is to write it to a file and then use your favorite viewing tool which you know works correctly with NLS to view the data.

Renita · Post by **Renita** » Wed Jul 02, 2008 2:01 am

Thank you for your prompt response. However, the map still does not give the output correctly. The 'O' character displayed with 2 dots on the top shows in the output, however the 'i' character with the dot shows up as '?'

I am using NLS ISO8859-9 and Turkish character set.

Please advice.

ArndW · Post by **ArndW** » Wed Jul 02, 2008 2:24 am

So if you read from SQL-Server and write to a sequential file the data has been corrupted? What character set is used in the SQL-Server data?

Renita · Post by **Renita** » Wed Jul 02, 2008 10:58 pm

Hi ArndW.

I am in the process of getting teh NLS setting from the SQL database admin.
However when I have just a database stage ii.e DRS stage and connect to the SQL database to recover data from a table, I see the o character with 2 dots but not the I with a single dot. I saw a post on IBM which said teh following

***Limitation on support for the Turkish locale.
For the WebSphere DataStage server canvas, a memory management problem occurs during casing operations when the CType locale category name is set to TR-TURKISH. Do not use this setting. For the WebSphere DataStage parallel canvas, locale support is available only for collation operations. Locale-based case operations are not supported for any locale. For casing operations in the Turkish locale, you cannot set the CType locale category name to tr_TR for parallel jobs. The special casing rules for Turkish letters dotted-I and dotless-I cannot be enforced.
********************************************************

However I am using server edition. I have set NLS to ISO8859-9 and turkish character set. Do you think I can have a work around this problem.

Appreciate your help on the same.

Thanks,
Renita

Renita · Post by **Renita** » Thu Jul 03, 2008 8:35 pm

Hi,

I got a reply back from the admin where they say that for their datastage edition 8.0 this issue got resolved using ASCL_EBCDIC-1026 character set. However I am on 7.5.2.

Since we are having an issue with just the I character in turkish, does anyone know if there is any fix available for the same.

Thanks,
Renita

Renita · Post by **Renita** » Thu Jul 03, 2008 8:36 pm

Hi,

I got a reply back from the admin where they say that for their datastage edition 8.0 this issue got resolved using ASCL_EBCDIC-1026 character set. However I am on 7.5.2.

Since we are having an issue with just the I character in turkish, does anyone know if there is any fix available for the same.

Thanks,
Renita